Question on Power BI Modeling View Referential Integrity and Normalization
Hi -
So, say we go to a company as a contractor to help them design a few views... I have a question for 2 scenarios for import vs direct queries.
- Say they give us a bunch of tables and we import the file, fine... do we need to normalize this data and check it for referential integrity.... Does this even matter in Power BI
- . Say we are working for an organization and the systems admins connect us to a SQL db and an Oracle db and some sales people give us a few tables.. maybe a salesforce connectivity... how do we know they normalized the data to match each other, before we start building... I "think" they have to talk and normalize before we connect first through the architect... that seems logical to me, but let's say we can't get that consensus, if the data is small in either of these situations than we need to ensure referencial integrity, clean it, vet it, simplify throw out stuff and then normalize it.. right? In theory... Right? Do I understand right?
1 answers ( 0 marked as helpful)
Power BI is a data visualization tool, it requires relational data so it can connect based on primary - foreign key relationships. The data needs to be normalized, you can use some of the PowerQuery functions to remove duplicate rows, specify how to handle null values, etc, but only if you know what conditions need to be met for each table, and things might get messier as you get more data. Power BI can handle data imports from various sources, but it needs to be normalized. Have you looked for a data warehouse?